J O U R N A L O F P R O T E O M I C S 1 0 3 ( 2 0 1 4 ) 1 7 8 – 1 9 3

Ava i l ab l e on l i ne a t www.sc i enced i r ec t . com

ScienceDirectwww.e l sev i e r . com/ loca te / j p ro t

Localization and proteomic characterization of

cholesterol-rich membrane microdomains in theinner ear

Paul V. Thomas, Andrew L. Cheng, Candice C. Colby, Liqian Liu, Chintan K. Patel,Lydia Josephs, R. Keith Duncan⁎

Kresge Hearing Research Institute, 5323 Medical Science Building I, 1150 West Medical Center Drive, The University of Michigan, Ann Arbor,MI 48109-5616, USA

A R T I C L E I N F O

⁎ Corresponding author. Tel.: +1 734 763 2129E-mail address: rkduncan@umich.edu (R

http://dx.doi.org/10.1016/j.jprot.2014.03.0371874-3919/© 2014 Elsevier B.V. All rights rese

A B S T R A C T

Article history:Received 10 February 2014Accepted 26 March 2014

Biological membranes organize and compartmentalize cell signaling into discretemicrodomains, a process that often involves stable, cholesterol-rich platforms that facilitateprotein–protein interactions. Polarized cells with distinct apical and basolateral cell processesrely on such compartmentalization to maintain proper function. In the cochlea, a variety ofhighly polarized sensory and non-sensory cells are responsible for the early stages of soundprocessing in the ear, yet little is known about the mechanisms that traffic and organizesignaling complexes within these cells. We sought to determine the prevalence, localization,and protein composition of cholesterol-rich lipid microdomains in the cochlea. Lipid raftcomponents, including the scaffolding protein caveolin and the ganglioside GM1, were found insensory, neural, and glial cells. Mass spectrometry of detergent-resistant membrane (DRM)fractions revealed over 600 putative raft proteins associated with subcellular localization,trafficking, and metabolism. Among the DRM constituents were several proteins involved inhuman forms of deafness including those involved in ion homeostasis, such as the potassiumchannel KCNQ1, the co-transporter SLC12A2, and gap junction proteins GJA1 and GJB6. Thepresence of caveolin in the cochlea and the abundance of proteins in cholesterol-rich DRMsuggest that lipid microdomains play a significant role in cochlear physiology.

Biological significanceAlthoughmechanismsunderlyingcholesterol synthesis, homeostasis, andcompartmentalizationin the ear are poorly understood, there are several lines of evidence indicating that cholesterolis a keymodulator of cochlear function. Depletion of cholesterol inmature sensory cells alterscalcium signaling, changes excitability during development, and affects the biomechanicalprocesses in outer hair cells that are responsible for hearing acuity. More recently, we haveestablished that the cholesterol-modulator beta-cyclodextrin is capable of inducing significantand permanent hearing loss when delivered subcutaneously at high doses. We hypothesizethat proteins involved in cochlear homeostasis and otopathology are partitioned intocholesterol-rich domains. The results of a large-scale proteomic analysis point to metabolicprocesses, scaffolding/trafficking, and ion homeostasis as particularly associated with

Keywords:HearingCochleaHair cellMicrodomainCaveolinLipid raft

; fax: +1 734 764 0014..K. Duncan).

rved.

179J O U R N A L O F P R O T E O M I C S 1 0 3 ( 2 0 1 4 ) 1 7 8 – 1 9 3

cholesterol microdomains. These data offer insight into the proteins and protein familiesthat may underlie cholesterol-mediated effects in sensory cell excitability and cyclodextrinototoxicity.

© 2014 Elsevier B.V. All rights reserved.

1. Introduction

The membrane lipid bilayer serves as a gatekeeper in cellphysiology, segregating biological reactions to specific extracellu-lar and intracellular spaces and serving as a hub of cell signaling.Fittingly, over 30% of the genes inmost genomes are estimated tocode for integral membrane proteins [1], and the proportion ofproteins associated with the membrane is even greater whenincluding those anchored by glycosylphosphatidylinositol (GPI)[2], fatty acid modifications [3,4], or other proteins. According tothe fluidmosaicmodel, thesemembrane proteins diffuse freelywithin a fluid lipid bilayer unless compartmentalized byprotein–protein interactions [5]. However, an accumulatingamount of evidence suggests that lipid–protein and lipid–lipidinteractions in the plasma membrane constrain the motion ofmembrane proteins and aid in their localization [6–8]. Theseobservations and others gave rise to the concept of lipid rafts,discrete 10–200 nm lipid-ordered microdomains enriched withsphingomyelin, glycosphingolipids, GM1 ganglioside, and cho-lesterol [6,7,9–12]. The partitioning of specific proteins withinthese microdomains appears to be a primary means forcompartmentalizing various processes in many, if not most,cell types. A large diversity in the structure and dynamics oflipid rafts [9,10,13] underlies their ability to organize a widearray of processes including protein and lipid trafficking,endocytosis, cellular recognition, and signaling [12].

The unique composition of lipid rafts renders theminsoluble in micelles of non-ionic detergents. These so-calleddetergent-resistant membrane (DRM) fragments can be sepa-rated from soluble membrane by sucrose density gradientfractionation [9,14]. Although DRM fragments and in situ lipidrafts are not to be equated, the analysis of DRM compositiontraditionally is the first step in identification of putative raftcomponents. Within the DRM, there are at least two domaintypes, those that incorporate the scaffolding protein caveolinand those that do not. Moreover, caveolin-based domains canbe further classified into two subtypes based on whetherthe scaffold assembles into planar rafts or morphologicallydistinct membrane invaginations termed caveolae [15].

In hair cells in the inner ear, observations of caveola-likestructures were reported decades ago [16,17], but little isknown about the role of cholesterol-enriched microdomainsin the inner ear. Hair cells contain diverse machinery, withmechanotransduction of sound energy occurring at the apicalend, synaptic transmission occurring at the basolateral end,and a complex network of ion channels functioning togetherto shape excitability. The polarity of these cells, their role insignal transduction, and the interplay of these ion channelsnecessitates the localization and co-localization of membraneproteins to specific regions of the cell. Multiple observationsincluding the presence of caveolae, inhomogeneous distribu-tion of membrane cholesterol [18,19], and detection ofextensive segregation of lipids in the membrane of the hair

bundle [20] suggest that the lipid bilayer of these cells servestomodulatemembrane protein distribution and functionality.While only caveolin and BK-type potassium channels havebeen biochemically identified in DRMs of cochlea [18], there isgrowing evidence that cholesterol-enriched microdomainsmay be involved in a wide range of processes in the innerear, including sensory transduction [20] and cochlear me-chanics [21–24]. Moreover, disruption of these microdomainswith cholesterol-chelating cyclodextrins causes aberrant elec-trophysiological behavior [18,22,23,25,26], while systemicdelivery of cyclodextrins causes profound hearing loss andouter hair cell death with no apparent effect on other systems[27].

To explore what processes may be affected by changes inmembrane cholesterol distribution in the inner ear, and themechanisms behind the inner ear's unique sensitivity, thelocalization and composition of cholesterol-enriched mem-brane microdomains were analyzed. The presence of raft-likedomains is supported by localization of the raft-associatedganglioside GM1 and the raft-scaffolding protein caveolin tosensory and non-sensory cells in the ear. Additionally, over600 proteins were found in triplicate preparations of cochlearDRM, and gene ontology analysis suggested DRM involvementin cell signaling, protein localization, and metabolism. Threegene products associated with syndromic hearing loss andeight gene products associated with non-syndromic hearingloss were also found in all three samples.

2. Methods

2.1. Tissue preparations

All animal procedures were conducted with the expressapproval of the University Animal Care and Use Committeeat the University of Michigan. White Leghorn chickens (Gallusgallus), 18–28 days old, were anesthetized with a ketamine/xylazine solution and euthanized by decapitation. Wholeauditory organs and microdissected subcompartments werecollected essentially as described previously [28]. Pectoralskeletal muscle, proventricular smooth muscle, heart, lung,and brain (cerebellum) were dissected rapidly on ice.

2.2. Gene expression assays

Total RNA was extracted from basilar papilla (10–12 persample) and auditory nerve (8–12 per sample) using anRNEasy Micro Kit (Qiagen), and from 50 to 200 mg samples ofall other organ tissues using Trizol reagent (Invitrogen) aftermechanical homogenization. RNA integrity was assessedusing an Agilent 2100 Bioanalyzer. Only samples with a28S:18S ratios greater than 1.0 were included in the analyses.First-strand cDNA was synthesized using Superscript III

180 J O U R N A L O F P R O T E O M I C S 1 0 3 ( 2 0 1 4 ) 1 7 8 – 1 9 3

Reverse Transcriptase (Invitrogen) according to themanufacturer's instructions.

Polymerase chain reaction (PCR) was used to verify thepresence of caveolin expression in the cochlea. Specific primersdesigned to amplify G. gallus caveolins included: Cav1 (GenBankaccession NM_0011056641) forward 5′-CAACATCTACAAGCCCAATAACAA-3′, reverse 5′-CTGAACACCTTGCCCATAGC-3′, amplicon438 bp; Cav2 (GenBank accession NM_001007086) forward5′-AAGAGCCTGACAGATGTTTTCGTT-3′, reverse 5′-AGCACAGTGAGGGCCAAAAATGAT-3′, amplicon 528 bp; Cav3 (GenBankaccession NM_204370) forward 5′-GTGGGGACGTACAGCTTTGATGGT-3′, reverse 5′-GAGGGAGTAGATGCGGCTGACACA-3′,amplicon 225 bp. PCR reactions were performed using PlatinumTaq DNA polymerase (Invitrogen) and 200 ng cDNA. PCRproducts were electrophoresed on a 1% agarose gel with 0.005%ethidium bromide. Bands were visualized with an AlphaInnotech FluorChem SP Imaging System (ProteinSimple). Theidentity of the amplicons was confirmed by DNA sequencing.

Real-time, quantitative polymerase chain reaction (qPCR)was conducted to determine differential expression of thecaveolin genes between tissues. These reactions were per-formed on an ABI Prism 7900HT Sequence Detection Systemusing custom-designed TaqMan probes and commercial PCRmaster mix (TaqMan 2× PCRMaster Mix, Applied Biosystems).Each sample was run in triplicate and was normalized to thehousekeeping gene S16 (Rps16). Primer combinations ampli-fied a single product of expected size, and amplificationefficiency for each gene product was within 90–100%. Primerand probe sequences for target genes included: Cav1 forward5′-ATGTCCGGCACCAAATACGT-3′, reverse 5′-GCTTGTAGATGTTGCCCTGTTC-3′, probe 5′-TCGGAGGGCTTTCTG-3′; Cav2 for-ward 5′-CCCCGCGGGCTGAA-3′, reverse 5′-CGGGCTCGGCGATCA-3′, probe 5′-CTCCAGCTGGGCTTC-3′; Cav3 forward5′-CCTGGTGAACAGAGATCCAAAGAG-3′, reverse 5′-CCACGGGCTCAGCTATCAC-3′, probe 5′-ATCCTCGAAATCCACCTTTAC-3′. Genes of interest were normalized to the referencegene Rps16 (S16, GenBank accession XM_416113), whichencodes a component of the small 40S ribosomal subunit[28]. Differences in expression level were calculated asfold-change relative to a control condition using the ΔΔCT

method [28,29].

2.3. Immunohistochemistry

Temporal bones were dissected and dura, oval window, andmiddle ear structures were removed. Bones were fixed in 2%paraformaldehyde (PFA) solution for 3 h at room temperaturethen decalcified for 2–3 days with 5% EDTA in 0.12 Mphosphate buffer (PB) with 0.2% PFA (pH 7.4–8.0) at roomtemperature. After rinsing with 0.5% sucrose in 0.12 M PB,tissues were cryoprotected via incubation with increasingconcentrations of sucrose (5%, 13%, 18%, 22%, and 30% in PB)for 30 min at room temperature and finally overnight at 4 °Cin 30% sucrose. Tissues were frozen in O.C.T. compound usingliquid nitrogen-cooled isopentane and stored at −80 °C. Thincryosections (10–12 μm) were collected alternately on approx-imately 20 slides so that any one slide contained representa-tive sections from 8 to 10 equally spaced regions along thelength of the cochlear duct. To label sections, slides werewarmed for 30 min at room temperature and washed in

phosphate buffered saline (PBS; Sigma D5652). Sections wereblocked in 5% normal goat serum and 1% bovine serumalbumin in PBS with 0.1% Triton X-100 (TX-100) for 1 h at roomtemperature. After removing the blocking buffer, tissuesections were exposed to primary antibody diluted in PBSplus 0.1% TX-100 overnight at 4 °C. Primary antibody bindingwas identified with AlexaFluor-conjugated goat anti-rabbit oranti-mouse secondary antibodies (Invitrogen) diluted 1:500 inPBS plus 0.1% TX-100, applied for 2 h at room temperature.Controls without primary antibody were prepared and ana-lyzed in parallel.

Dissociated single hair cells were prepared for immunocy-tochemistry to investigate subcellular compartmentalizationof microdomain structures. To dissociate single cells, basilarpapillae were treated for less than 1 min with 0.01% proteasetype XXIV (Sigma, P8038) to facilitate removal of the tegmen-tum vasculosum and tectorial membrane. A borosilicatecapillary tube was used to aspirate the sensory epitheliumand mechanically dissociate the cells onto a slide containing(in mM) 150 NaCl, 6 KCl, 5 CaCl2, 2 MgCl2, and 5 HEPES,buffered to pH 7.4 with NaOH. Cells were allowed to settle onSuperfrostPlus slides for 10–15 min to promote adherence,then fixed with 2% PFA for 30 min at room temperature. Slidesof dissociated cells were blocked and labeled as above, withthe exception of using 0.01% saponin in place of 0.1% TX-100when labeling fixed hair cells with anti-Cav1.

Primary antibodies included rabbit anti-caveolin (BD Trans-duction Laboratories, 610059), mouse anti-Cav1 (BD Transduc-tion Laboratories, 610406), mouse anti-Cav2 (BD TransductionLaboratories, 610684), mouse anti-Cav3 (BD TransductionLaboratories, 610420), and rabbit anti-neurofilament M 145–kD(Chemicon, AB1987). In some cases, hair cell preparations weretreated with 1:200 OG-phalloidin (Invitrogen, O7466) to identifyactin-rich hair bundles. When required, nuclei were counter-stained with 1:50 Hoechst 33342 (Invitrogen, H3570). Sectionsand cells were imaged on a Leica DM LB fluorescencemicroscope (Leica Microsystems). Images were acquired usingsingle- and triple-band filter sets and a cooled-CCD color digitalcamera (MicroPublisher, QImaging).

2.4. Isolation of detergent resistant membrane fractions

Twelve whole cochlear ducts per sample were extracted, pooledinto a single tube, and flash frozen on dry ice. Samples werethawed on ice in 120 μl of MES buffered saline (MBS; 25 mMMES,150 mM NaCl, pH 6.5) containing 0.25%, 0.5%, 1%, or 2% (v/v)TX-100 and protease inhibitor cocktail (Thermo Fisher, PI-78410).After homogenization with a motorized pestle, lysates wereincubated on ice for 30 min. The homogenate was placed on thebottom of a 2.2 ml ultracentrifuge tube andmixed 1:3 with 53.3%(w/v) sucrose/MBS to a final concentration of 40% (w/v) sucrose/MBS. For a discontinuous density gradient, this solution wasoverlaid with 900 μl of 30% sucrose/MBS, then overlaid with1 mL of 5% sucrose/MBS and balanced. Membrane proteinswere separated using an Optima Max-E Ultracentrifuge witha swinging bucket rotor (TLS-55) (Beckman-Coulter, Fuller-ton, CA). The samples were ultracentrifuged at 54,000 rpm(~200,000 g) for 24 h at 4 °C. Twelve equal volume fractions(183 μl each) and a pellet fraction were collected from thetop.

181J O U R N A L O F P R O T E O M I C S 1 0 3 ( 2 0 1 4 ) 1 7 8 – 1 9 3

2.5. Immunoblot

Aliquots (25 μl) were taken from each fraction, mixed withLaemmli sample buffer (Bio-Rad 161-0737) containing 5%2-mercaptoethanol, separated by SDS-PAGE on a 4–15% poly-acrylamide gel, and transferred to nitrocellulose membranes(Pierce Biotechnology, Rockford, IL) for Western blotting.Membranes were probed with a primary antibody to caveolin(1:1000 anti-caveolin, BD Transduction Laboratories, 610059)or transferrin receptor (1:1000 anti-TfR, Invitrogen, 13-6800).Secondary antibodies included goat anti-mouse or anti-rabbitconjugated to horseradish peroxidase (1:10,000, Thermo Scien-tific Pierce, 31432 or 31460). Reactions were visualized withSuperSignal West Femto chemiluminescent substrate (ThermoFisher Pierce, 34095) and imaged on an Alpha InnotechFluorChem SP Imaging System (ProteinSimple).

2.6. Labeling of GM1-enriched domains

The localization of GM1-enriched domains was visualized withfluorescently conjugated cholera toxin subunit B (CTX) (FITC-CTX, Sigma C1655; Alexa Fluor 488-CTX, Invitrogen C34775).Dissociated cochlear hair cells were pre-fixed with 2% PFA for30 min, treated with 0.01% saponin for 15 min, and exposed to8–20 μg/ml CTX for 30 min at room temperature. In some cases,CTX was directly applied to freshly dissociated hair cells in theabsence of PFA and saponin. Punctate aggregates of CTX labelwere counted on pre-fixed, 488-CTX-labeled cells. Clearlyisolated puncta greater than 0.3 μm in diameter and locatedwithin the outline of the basolateral cell (i.e. excluding the hairbundle and cuticular plate) were included in these counts.

Dot blots of subcellular membrane fractions were labeledwith HRP-conjugated CTX (Invitrogen, C34780). One microliterof each fraction was applied to strips of dry nitrocellulosemembrane. The strips were blocked for 1 h at room temper-ature in 5% (w/v) non-fat milk in TBS containing 0.1%(v/v)Tween (TBS-T) and incubated at 4 °C overnight in 200 ng/mLHRP-CTX in 5% non-fat milk in TBS-T. Reactions werevisualized with SuperSignal West Femto chemiluminescentsubstrate and imaged on an Alpha Innotech FluorChem SPImaging System.

2.7. Protein quantification

The micro bicinchoninic acid (BCA) protein assay (PierceBiotechnology, 23235) was used for protein quantification.Fractions were diluted with distilled water to reach proteinconcentrations within the linear range of the assay. Absor-bances were measured at 590 nm using an automatedmicroplate reader (EL311, Bio-Tek Instruments).

2.8. Cholesterol quantification

The Amplex Red cholesterol assay (Invitrogen, A12216)was used for cholesterol quantification according to themanufacturer's instructions, and fluorescence was measuredusing a Fluoroskan plate reader (Thermo Fisher ScientificInc.). Fractions were diluted with the provided reaction bufferto reach cholesterol concentrations within the linear range ofthe assay.

2.9. Mass spectrometry

Detergent-resistant membrane fractions from three indepen-dent biological samples were submitted to MS Bioworks forproteinmass spectrometry analysis. Following sucrose-gradientseparation, fractions 5–7 for a given sample were pooled anddissolved in MBS with protease inhibitor cocktail (ThermoFisher, PI-78410) for a final volume of 2.3 mL. The dilutedsamples were centrifuged for 1 h at 54,000 rpm (~200,000 g) at4 °C. The pellet was resolubilized in 100 μL of 2× LaemmliSample Buffer (Bio-Rad, 161-0737) diluted 1:1 in MBS without2-mercaptoethanol. Protein concentrations were measured and20 μg of protein from each sample was separated ~1.5 cm on a10% Bis-Tris Novex mini-gel (Invitrogen) using the MES buffersystem. The gel was stained with Coomassie blue and each lanewas excised into ten equally sized segments. Using a robot(ProGest, DigiLab), gel pieces were processed by being washedwith 20 mM ammonium bicarbonate followed by acetonitrile,then reduced with 10 mM dithiothreitol at 60 °C, followed byalkylation with 50 mM iodoacetamide at room temperature,digested with trypsin (Promega) at 37 °C for 4 h, and quenchedwith formic acid. The supernatant was analyzed directlywithout further processing. Each fraction was analyzed bynano LC/MS/MS with a Waters NanoAcquity HPLC systeminterfaced to a ThermoFisher LTQ Orbitrap Velos Pro. Peptideswere loaded on a trapping column and eluted over a 75 μmanalytical column at 300 nL/min; both columns were packedwith Jupiter Proteo resin (Phenomenex). The mass spectrome-ter was operated in data-dependent mode, with MS performedin the Orbitrap at 60,000 FWHM resolution and MS/MSperformed in the LTQ. The fifteen most abundant ions wereselected for MS/MS. Spectral data were searched using Mascot(Matrix Science) with the following parameters: enzyme,trypsin; Database, Uniprot Chicken (concatenated forwardand reverse plus common contaminants); fixed modification,carbamidomethyl (C); variable modifications, oxidation (M),acetyl (N-term), Pyro-Glu (N-term Q), deamidation (N,Q); massvalues, monoisotopic; peptide mass tolerance, 10 ppm; frag-ment mass tolerance, 0.8 Da; Max missed cleavages, 2.

Mascot DAT files were parsed into Scaffold4 (ProteomeSoftware) for validation, filtering, and to create a non-redundant list per sample. Data were filtered using aminimum protein value of 99.0%, a minimum peptide valueof 50.0% (Prophet scores), and a minimum of two uniquepeptides per protein. Protein and peptide false discovery rateswere both set to 1%. The spectral abundance factor for eachprotein was calculated by dividing the number of spectralcounts of assigned peptides by the protein molecular weightin kDa. The spectral abundance factor for each protein wasdivided by the sum of all of the spectral abundance factors ineach biological sample, disregarding contaminants and decoyproteins, resulting in the normalized spectral abundancefactor (NSAF) [30,31]. The NSAF for each protein was averagedover all three biological samples.

2.10. Gene ontology and prediction tools

Gene ontology (GO) analysis was conducted using blast2GO[32–35] and the ClueGO extension [36] for Cytoscape 3.0.1 [37].Protein FASTA sequences for each UniProt entry were entered

182 J O U R N A L O F P R O T E O M I C S 1 0 3 ( 2 0 1 4 ) 1 7 8 – 1 9 3

into blast2GO and all default settings for blast2GO were used,except when creating the combined graph for biologicalprocess GO terms in which a minimum of 3 sequences wasrequired to create a node. All annotation tools in blast2GOwere used, except GO-slim. Gene symbols inferred fromhomology were submitted into ClueGO and all default ClueGOsettings were used, except for the following: GO terms werelimited to one level at a time, and the minimum numberof sequences per node was set to 1. GO term enrichmenttests were conducted according to hypergeometric tests andcorrected using Bonferroni step down analysis [36,38].

Transmembrane helix prediction was conducted using theonline TMHMM 2.0 prediction server (http://www.cbs.dtu.dk/services/TMHMM/) [1,39] and palmitoylation was assessedusing CSS-Palm 3.0 [40,41]. Glypiation was predicted using theonline predGPI prediction server (http://gpcr.biocomp.unibo.it/predgpi/) [2], and myristoylation was predicted using theonline Expasy Myristoylator prediction server (http://web.expasy.org/myristoylator/) [42].

3. Results

Dissociated hair cells were stained by fluorescently-conjugatedcholera toxin β-subunit (CTX), a marker for GM1 gangliosidesand lipid rafts. Live-cell preparations exhibited punctate CTX

Fig. 1 – CTX labeling of isolated hair cells reveals punctate GM1-hair cells from throughout the abneural and neural regions of thapplication. Unlabeled preparations and those treated with uncoLive cells displayed punctate spots between 0.3 and 1 μm in diabasolateral surface (arrowhead). Punctate spots were exclusivelyoutlined with thin lines to show cell orientation. Scale bar = 10 μ

spots at subnuclear and supranuclear ends of the hair cell(Fig. 1A). Exemplar cells classified as “short”, “intermediate”,and “tall” based on length/width ratios were similarly labeled.These morphologies reflect distinct functional roles in thechick cochlea [43]. The results, therefore, indicate that CTXshows no preference for hair cell subtype. More recently, CTXhas been applied to fixed preparations in order to limit thepossibility of CTX-induced endocytosis and translocation ofGM1-rich membrane domains [44,45]. Labeling in fixed haircells resulted in more numerous, smaller, and more distinctpuncta located primarily in subnuclear regions of the cell(Fig. 1B). Approximately 60% of isolated, fixed hair cellsdemonstrated quantifiable punctate spots. On average, labeledhair cells exhibited 3–4 puncta (Fig. 2). The remaining cellsincluded those that were unlabeled, those masked by debris,and those with indistinct puncta (i.e. too small or without clearboundaries).

While GM1 is often found co-localized with the cholesterol-binding caveolin, there is evidence in chick brain that GM1 andcaveolin are members of separate populations of DRM [46]. Toprobe caveolin expression, total RNA was collected from wholecochlea, auditory nerve, skeletal muscle, cerebellum, andsmooth muscle and probed for the expression of caveolin 1–3(Fig. 3A). All three members of the caveolin gene family weredetected in whole cochlea and auditory nerve. The presence ofcaveolin 3 in the cochlea was somewhat surprising since this

rich microdomains. Unfixed (A) and pre-fixed (B), dissociatede cochlea exhibited bright, punctate spots following CTXnjugated CTX were devoid of punctate staining (not shown).meter (arrow) as well as large, diffuse domains along thefound on pre-fixed, CTX-labeled cells. Hair bundles arem.

Fig. 2 – Histogram of CTX puncta shows an average of 3–4puncta per hair cell. Pre-fixed, isolated hair cells displayingone or more CTX-puncta were analyzed, revealing a mean of3.8 ± 0.3 (standard error of the mean) puncta per cell. N = 40.

183J O U R N A L O F P R O T E O M I C S 1 0 3 ( 2 0 1 4 ) 1 7 8 – 1 9 3

product is primarily associated with muscle and glia. To probefurther, qPCR was used to examine caveolin expression inthe microdissected, hair cell-enriched sensory epithelium andcompared to other tissues (Fig. 3B–D). The auditory sensory

Fig. 3 – All three members of the caveolin gene family are expresnerve. (A) RT-PCR of whole cochlea, auditory nerve, and select coCav3. Quantitative PCR shows the differential expression of (B) Canerve (AN), skeletal (SkM) and smooth (SmM) muscle, heart (H), aOne-way ANOVA with post-hoc pairwise comparison to Br.

epithelium and lung were significantly enriched in caveolin 1compared to brain (P < 0.05, one-way ANOVA), whereas cave-olin 2 expression was more uniform between tissues. Skeletalmuscle and heart were significantly enriched in caveolin3 (P < 0.05, one-way ANOVA), while the other tissues hadrelatively similar levels compared to brain.

To localize the expression of caveolin in the inner ear,sections of chicken cochlea and dissociated hair cells wereprobed for the presence of caveolin 1 (Fig. 4) and caveolin 3(Fig. 5); antibodies to caveolin 2 revealed no specific stainingabove background (data not shown). Radial frozen sectionsof the basilar papilla stained with anti-caveolin 1, showingbright label in hair cells throughout the neural (left) toabneural (right) cross-section (i.e. no preference for short ortall hair cells) (Fig. 4A–B). Primary auditory neurons and theirperipheral projections were strongly labeled by anti-caveolin1 (Fig. 4C). Similar to results with CTX, approximately 50% ofisolated hair cells revealed several punctate caveolin-1-positive spots (Fig. 4D). Anti-caveolin 3 labeled throughoutthe sensory epithelium making it difficult to distinguishbetween specific labels in hair cells, supporting cells, andneuronal cell types (data not shown). The strongest caveolin 3staining was found in regions surrounding the soma ofprimary auditory neurons (Fig. 5A). To confirm that this labelwas associated with small diameter glial cells associated with

sed in the chick cochlear sensory epithelium and auditoryntrol tissues reveal bands corresponding to Cav1, Cav2, andv1, (C) Cav2, and (D) Cav3 in sensory epithelium (SE), auditorynd liver (L) tissue relative to cerebellum (Br). *P < 0.05,

Fig. 4 – Anti-caveolin 1 labels hair cells and auditory neuronsin chick cochlea. (A) A radial section from the distal(low-frequency) portion of the chick basilar papilla is shownto illustrate the position of hair cells (HC) and supportingcells (SC) of the sensory epithelium, neighboring homogenecells (HG), and the tectorial membrane (TM). (B) Anti-caveolin1 brightly labeled the basolateral portion of the hair cells.As is often the case, the tectorial membrane wasnon-specifically reactive with the secondary antibody.Supporting cells were not immunoreactive. (C) Primaryauditory neurons were specifically labeled by anti-caveolin1. (D1-4) Isolated hair cells were diffusely labeled byanti-caveolin 1, with approximately 50% of the cells alsodisplaying punctate basolateral spots (among 20 imagedhair cells). Hair bundles (all pointing upward) were notimmunoreactive. Scale bars, A–C = 10 μm, D = 20 μm.

184 J O U R N A L O F P R O T E O M I C S 1 0 3 ( 2 0 1 4 ) 1 7 8 – 1 9 3

these neurons, teased nerve preparations were stained forcaveolin 3 and neurofilament. In some cases, large diameterneurofilament-positive neurons were found with tightlybound, small-diameter, neurofilament-negative/caveolin-3-positive cells indicative of caveolin 3 expression in glial cells(Fig. 5B). Approximately 30% of isolated hair cells showedevidence of punctate caveolin-3-positive spots at the basolateralpole of the cell (Fig. 5C).

To begin unraveling the composition of caveolin-basedand GM1-rich microdomains, whole cochlear lysates weretreated with cold TX-100 and subjected to sucrose-gradientfractionation. Using 1% TX-100, the composition of proteins inDRM fractions differed from those in detergent-sensitive andpellet fractions (Fig. 6). While this concentration of TX-100 is a

useful starting point separating DRM and non-DRM, someoptimization was required to (1) fully solubilize non-DRMwhile leaving DRMs intact, (2) identify those fractions thatcontain both caveolin and GM1, and (3) confirm the enrich-ment of cholesterol in those DRM/caveolin/GM1 fractions.High detergent concentrations (1% [Fig. 7] and 2% TX-100 [notshown]) resulted in a broad distribution of caveolin and GM1across low and high numbered fractions, overlapping with thenon-DRM marker transferrin receptor. Similarly, at 0.25%TX-100, caveolin monomers and putative dimers were dis-tributed across most fractions, overlapping with transferrinreceptor. The separation was optimized with 0.5% TX-100with faint overlap of caveolin/GM1 and transferrin receptor infraction 8. Under these conditions, caveolin and GM1 wereconcentrated particularly in fractions 5, 6, and 7.

Likewise, bulk protein partitioning in DRM fractions was bestseparated from soluble fractions with 0.5% TX-100. The distri-bution of protein in fractions 4–11 was reliably dependenton TX-100 concentration (Fig. 8) (P < 0.05, two-way ANOVA). For0.25% TX-100, total protein was undetectable in the lowestdensity fractions (1–4), peaking in fraction 8 and reducing onlyslightly before rising again in fractions 11–12. In higherconcentrations of TX-100, protein was first detectable infractions 3 and 4. Treatment with 1% TX-100 resulted ina gradually increasing distribution of protein from low-to-high-density fractions; there was no apparent bimodalseparation of protein in DRM and non-DRM fractions. The 0.5%TX-100 condition, however, showed a distinct peak of proteinconcentration in fraction 6, consistent with the separation ofcaveolin and transferrin receptor seen in immunoblots. Like-wise, cholesterol was concentrated in fractions 5–7 under theseconditions, peaking in fraction 6, whereas higher numbered,detergent-soluble fractions were depleted of cholesterol (Fig. 9).Together, these data depict an optimized separation betweenDRM and non-DRM pools using 0.5% TX-100.

LC-MS/MS of pooled DRM fractions 5, 6, and 7 identified 967proteins out of three separate biological samples, with 14contaminants and 9 decoy proteins using the criteria de-scribed (Supplementary Table 1). A total of 606 proteins werecommon to all three separate biological samples and wereused for all further analyses. Using gene ontology (GO)analysis, twenty-six (4%) of these DRM proteins were well-characterized raft-related proteins, including caveolin-1,thy-1, and flotillin-2 (Table 1). Only 264 (44%) of the identifiedproteins had predicted transmembrane helices based onprotein sequence (Supplementary Table 1). Protein sequenceswere examined for post-translational modifications that cantarget peripheral membrane proteins to DRM (glypiation,palmitoylation, and myristoylation) [12,47] using predictionalgorithms. The presence of sites specific to each of these threepost-translational modifications and association with proteincomplexes are indicated in Supplementary Table 1 andsummarized in Table 2. In total, 20 proteins were predicted tobemyristoylated, and 26were predicted to be glypiated with nooverlap. In contrast, 462 proteins (76%) had predictedpalmitoylation sites. Additionally, using GO analysis, 247proteins (41%) were identified as well-characterized constitu-ents of protein complexes, which could facilitate membraneassociation by incorporating integral membrane componentsinto the complex. Importantly, there is significant overlap

Fig. 5 – Anti-caveolin 3 labels hair cells and non-neuronal cells in the auditory ganglion. A cryosection of the auditory ganglionis shown stained for caveolin 3 (A1) and neurofilament (A2), revealing caveolin staining in glia surrounding the neuron cellbodies. Teased nerve preparations of isolated neuron/glia complexes (B1) were stained with anti-caveolin 3 (red),neurofilament (green), and Hoechst (blue) and imaged with a triple-band filter (B2). Neurofilament-negative cells labeledstrongly with anti-caveolin 3. Isolated hair cells were labeled throughout by anti-caveolin 3, except the cuticular plate at thebase of the hair bundle (C1–2). Approximately 30% of the cells displayed punctate spots at the base of the cell (C2, among 20imaged hair cells). Scale bars, A–B = 20 μM, C = 10 μM.

185J O U R N A L O F P R O T E O M I C S 1 0 3 ( 2 0 1 4 ) 1 7 8 – 1 9 3

between proteins with predicted palmitoylation sites and allother categories. Only 41 proteins (7%) were not predicted to beassociated with the membrane by any of these methods.

The 20 DRM proteins with highest NSAF are described inTable 3. Peptide spectra matched to these proteins accountedfor 23.7% of the NSAF. The majority of these proteins areinvolved in mitochondrial metabolic processes includingproteins from the respiratory chain, namely ATP synthaseand cytochrome C oxidase components, and proteins in-volved in transporting organic molecules across the mito-chondrial outer membrane. Well-known constituents of lipidmicrodomains in the plasma membrane and intracellularorganelles, such as prohibitins, annexins and myelin proteinzero were also abundant. GO analysis of all 606 DRM proteinsby biological process revealed top level 2 categories (>5% ofNSAF) of cellular processes, metabolic processes, localizationand organization, biological regulation and response tostimulus, multicellular processes and signaling (Fig. 10A).Top level 2 categories of cellular component (>1% of NSAF) GOanalysis were cell, membrane, protein complexes, and organ-elles, specifically membrane-bounded organelles (Fig. 10B).

While the GO categories of DRM proteins illustratethe locations and processes in which cholesterol-enriched

membrane microdomains may play a role, by focusing onthe total spectral abundance of proteins associated with aspecific GO category, this approach gives more importance toGO categories that have more database entries. Additionally,this method limits the utility of more specific, higherresolution GO categorization by providing an unmanageablylarge number of distinct, but related GO terms. Fig. 11 depictssignificantly enriched (P < 0.01), level 3 GO categories bycomparing the number of genes that are associated witheach GO category with the total number of genes associatedwith that category in the GO database. Using biologicalprocesses GO terms, most of the top categories in Fig. 10A arestill represented, but now, other important GO categories areincluded such as those related to various transport processesand subcellular localization (Fig. 11A). This approach to GOanalysis highlights many smaller and more specific cellularcomponent GO categories (Fig. 11B), including those associ-ated with cell–cell junctions and protein anchoring at theplasma membrane. The membrane raft component was alsosignificantly enriched (P < 0.05) (not shown).

By far, processes associated with metabolism and bioener-getics involved the majority of identified DRM proteins (63.9%),which carries particular importance given the cochlea's acute

Fig. 7 – The separation of DRM and non-DRM membranefractions is dependent on TX-100 concentration. Westernblots and dot blots of fractionated whole cochlear lysates areshown following treatment with 1%, 0.5%, or 0.25% TX-100.Western blots were probed with pan-caveolin (Cav) andtransferrin receptor (TfR) antibodies to identify DRM andnon-DRM fractions, respectively. Dot blots were probed withHRP-conjugated CTX to identify GM1-enriched fractions.

Fig. 8 – Protein concentration among DRM and non-DRMfractions varies considerably with concentration of TX-100.Protein concentration was quantified by BCA assay andaveraged for fractions 1–11 of 0.25%, 0.5%, and 1% TX-100treated whole cochlea lysates. N = 3 for each group. Errorsrepresent one standard deviation from the mean.

Fig. 6 – SDS-PAGE of cochlear fractions revealsheterogeneous protein distributions among DRM, non-DRM,and soluble protein pools. Whole cochlea lysates treatedwith 1% TX-100 were separated by SDS-PAGE and stainedusing Coomassie Blue to reveal the diversity and abundanceof proteins in pooled DRM fractions (4–6), pooled detergentsensitive fractions (9–11), and the pellet of soluble proteins.

186 J O U R N A L O F P R O T E O M I C S 1 0 3 ( 2 0 1 4 ) 1 7 8 – 1 9 3

sensitivity to mitochondrial dysfunction. However, the analy-ses in Fig. 11 highlight additional processes/components thatplay influential roles in the cochlea, including those involved inion homeostasis, exocytosis and vesicular transport, and cell–

Fig. 9 – Cholesterol is specifically associated with DRMfractions. Cholesterol concentration was determined usingAmplex Red enzyme assays for fractions 1–12 of 0.5% TX-100treated whole cochlea lysates and averaged across 3independent samples. Error bars represent one standarddeviation from the mean.

Table 3 – Identified DRM proteins with highest NSAF.

Genename

Gene description NSAF

VDAC2 Outer mitochondrial membrane porin 2(voltage dependent)

2.59%

COX6C Cytochrome C oxidase, subunit 6C 2.02%MPZ Myelin protein zero 1.91%SLC25A4 Mitrochondrial ADP/ATP translocase 1 1.55%ATP1A1 ATPase, Na+/K+ transporting, alpha 1

polypeptide1.52%

COX4I1 Cytochrome C oxidase subunit 4, isoform 1 1.20%ANXA2 Annexin A2 1.16%ATP5A1 ATP synthase, H+ transporting, F1 complex,

alpha subunit 11.12%

ATP5B ATP synthase, H+ transporting, F1 complex,beta subunit

1.08%

CISD1 CDGSH iron sulfur domain 1 1.02%ACTG1 Cytoplasmic/cytoskeletal actin, gamma 1 0.99%COX6A1 Cytochrome C oxidase, subunit 6A 0.95%SLC25A6 Mitochondrial ADP/ATP translocase 3 0.87%VDAC1 Outer mitochondrial membrane porin 1

(voltage dependent)0.86%

CKAP4 Cytoskeleton-associated protein 4 0.84%SLC25A3 Mitochondrial phosphate carrier protein 0.83%PHB Prohibitin 0.82%PHB2 Prohibitin 2 0.80%ATP5H ATP synthase, H+ transporting, F0 complex,

subunit D0.80%

ANXA6 Annexin A6 0.79%

Table 1 – Identified DRM proteins associated with the GOterm “membrane raft”.

Gene description Gene name

ATPase, Na+/K+ transporting, alpha 2polypeptide

ATP1A2

ATPase, Ca2+ transporting, plasmamembrane 1

ATP2B1

Caveolin 1 CAV1Cadherin 13 CDH13Dystrophin DMDFlotillin 2 FLOT2Connexin 43 GJA1G-protein subunits/polypeptides GNA11, GNAI2,

GNAI3, GNASGlypican 1 GPC1Inositol 1,4,5-triphosphate receptors ITPR1, ITPR3K-Ras P21 protein KRASLate endosomal/lysosomal adaptor andMAPK and MTOR activator 3

LAMTOR3

Lck/Yes-related novel protein tyrosine kinase LYNNeimann–Pick C1 protein NPC1Protein kinase C, alpha type PRKCAMajor prion protein PRNPRAS-related protein Rab-5A RAB5AHEPG2 glucose transporter type 1,erythrocyte/brain

SLC2A1

Syntaxins STX2, STX12Thy-1 membrane glycoprotein/surfaceantigen

THY1

187J O U R N A L O F P R O T E O M I C S 1 0 3 ( 2 0 1 4 ) 1 7 8 – 1 9 3

cell interactions befitting an epithelial tissue separated byunique ionic environments. To further explore the DRMcomponents involved in these processes, we examined thoseproteins lying at the key intersections of vesicular transportthrough which the trafficking of proteins and synaptic signal-ing in hair cells occur (Supplementary Table 3), ion transportthroughwhich biological regulation, signaling and homeostasisoccur (Supplementary Table 4), and cell junctions wheresignaling and protein localization are crucial (SupplementaryTable 5). Over 100 proteins were associated with the GO terms“vesicle” and/or “vesicle-mediated transport” (SupplementaryTable 3). Thirty-one were associated with the Golgi orendoplasmic reticulum membrane, where proteins may beembedded in nascent lipid microdomains. Interestingly, 19 ofthese vesicle-associated proteins were associated with thesynapse, and 42 were associated with exocytosis/endocytosis,essential processes in normal cochlear physiology. Supple-mentary Table 4 depicts a complex set of sodium, potassium,calcium, chloride and magnesium transporters and channels.Over 100 proteins were associated with ion transport or ion

Table 2 – Predicted transmembrane helices, post-translational mDRM proteins. Percentage of proteins associated with only a sinoff-diagonal. Combined percentage of all proteins in a category

Transmembrane Palmitoylation Glypiation M

Transmembrane 4.6% 34.7% 1.7%Palmitoylation 19.6% 4.0%Glypiation 0.2%MyristoylationProtein complex

flux, suggesting that DRM microdomains may be essentialcomponents of ion homeostasis and energy metabolism in theear. Finally, the GO terms “cell junction” and “cell junctionorganization” were combined to examine DRM proteins thatmay be involved in the conservation of the endocochlearpotential and potassium recycling in the inner ear [48](Supplementary Table 5). Key cell junction proteins, includingconnexins, catenins, plakophilins, spectrins, and claudin-1,were present in inner ear DRM along with a number of celladhesion related surface proteins, such as CD9 and thy-1.These results suggest that cholesterol-rich domains signifi-cantly contribute to the tight-junction and gap-junctionnetworks involved in potassium recycling within the ear.

Among the proteins found in DRM microdomains, eightwere associated with non-syndromic deafness and three withsyndromic deafness (Table 4). These proteins are primarilyinvolved in ion transport and homeostasis, cell–cell junctions,vesicular transport, and extracellular structures, broadlyreflecting the range of enriched components/processes iden-tified by the GO analysis.

odifications, and protein complex association of identifiedgle category on diagonal and those with pairs of features onshown in bold.

yristoylation Protein complex Combined for each category

0.2% 15.2% 43.6%2.1% 17.5% 76.2%0% 1.0% 4.3%

0.8% 0.3% 3.3%6.8% 40.8%

Fig. 10 – GO associations are listed for identified DRM proteins. Annotated proteins were weighted by the normalized spectralabundance factor and analyzed using blast2GO to identify major GO associations. (A) Level 2 biological process GO termsare listed for terms with a total NSAF above 5%. (B) Level 3 cellular component GO terms are listed for terms with a total NSAFabove 1%.

188 J O U R N A L O F P R O T E O M I C S 1 0 3 ( 2 0 1 4 ) 1 7 8 – 1 9 3

4. Discussion

Mounting evidence supports a role for cholesterol in thesegregation of membrane domains in the cochlea and inmodulating cochlear physiology [18,20,23,25,26]. Until now,such studies have only hinted at the involvement of raft-like,lipid-orderedmicrodomains that, in other systems, limit lateraldiffusion of membrane-associated proteins and aid in thecompartmentalization of cell signaling. Our study representsthe first comprehensive analysis of the localization andcomposition of cholesterol-enrichedmembrane microdomainsin a peripheral auditory organ. The raft markers caveolin andGM1 were found in both sensory and non-sensory cells in theinner ear, and a screen of cochlear DRM fractions revealedhundreds of proteins that co-segregated with cholesteroland GM1. Bioinformatic analysis of these microdomainsidentified proteins involved in metabolic, trafficking, andsignaling processes that are essential to normal cochlearfunction.

Caveolin is a scaffolding protein that identifies a subclassof lipid raft, structurally supporting either planar domains orinvaginating caveolae. Caveolin expression was relativelyhigh in the cochlea and auditory ganglion, compared toother central and peripheral tissues. Immunohistochemistrydetected Cav1 and Cav3 in sensory cells, Cav1 in auditoryneurons, and Cav3 in glia. Caveolin-2 antibodies failed toreveal specific staining patterns, but this may be attributed tolimited cross-reactivity with chick isoforms. Based on the

expression of Cav2 mRNA in cochlea and similarities in thetissue distribution of Cav1 and Cav2 in other systems [49] it isquite possible that Cav2 also plays a role in inner ear whereverCav1 is found. Caveolins are widely distributed in adipocytes,epithelial, and endothelial cells, as well as sensory, neural,andmuscle cells [50–53]. Caveolin interacting proteins includeSrc family tyrosine kinases, nitric oxide synthase, receptortyrosine kinases, phospholipase C, protein kinase C, Rasproteins, and G protein subunits [51]. Many of these proteinsplay important roles in cochlear signaling, development ofhearing, and cochlear response to injury.

Caveolin and GM1-rich microdomains were identified inisolated hair cells as discrete puncta located along thebasolateral membrane. Unfortunately, the size and subcellularlocalization of raft domains can be confounded by coalescenceand redistribution during labeling. For example, when appliedto live cells, CTX can induce the internalization of GM1-richvesicles, whereas application of CTX to lightly fixed, coldpreparations can limit redistribution of GM1 [45]. Theseobservations may explain the appearance of large and morebroadly distributed CTX-domains in live compared to fixed haircells. Taken together, our data would suggest that GM1-richdomains are present at the synaptic pole of the cell, consistentwith caveolin-1 and caveolin-3 staining patterns. However, itremains unclear whether these markers labeled the samestructures [46,54]. Nevertheless, it is tempting to associatethese basolateral microdomains with synaptic active zones,calcium hotspots, and BK-channel puncta that all exist alongthe basal hair cell membrane and are similar to CTX-positive

Fig. 11 – Significantly enriched GO categories are listed for identified DRM proteins. (A) Level 3 biological process GO terms and(B) Level 3 cellular component terms with significant enrichment (P < 0.01). Terms are scored according to the number of genesidentified in the DRM that are associated with a given term as a percent of the total number of genes associated with that term.

189J O U R N A L O F P R O T E O M I C S 1 0 3 ( 2 0 1 4 ) 1 7 8 – 1 9 3

spots in both size and number [55–57]. This inferenceis supported by evidence that BK-channel subunits areincorporated into DRM and that both BK activity and calciumchannel function are modulated by the DRM-disruption agentmethyl-beta-cyclodextrin [18]. Further study is required todefine the roles of GM1 and caveolin in hair cells and to testwhether these domains organize channel clusters and synapticmachinery.

As a first step toward identifying the protein composition oflipid rafts in the cochlea, we optimized the separation of DRMfrom non-DRM fractions by titrating the concentration ofTX-100. Although 1.0% TX-100 has been considered a gold-

standard for defining DRM, protein recovery in DRM fractionsis exquisitely sensitive to detergent:protein ratio [58,59]. Inchick cochlea under these homogenization conditions, a lowerconcentration of 0.5% TX-100 yielded optimal separation,reducing loss of DRM proteins from over- or under-solubilization. When subjected to mass spectrometry, over 960proteins were identified in three separate cochlear DRMpreparations. About 600 of these proteins were common toeach biological repeat with the remainder largely representinglow abundant proteins at the detection limits of our approach.We identified a relatively large set of DRM proteins compared tosimilar studies, which reported 70 DRM proteins in Jurkat T cells

Table 4 – Identified DRM proteins associated withdeafness.

Genename

Gene description Deafness

ACTG1 Cytoplasmic/cytoskeletal actin,gamma 1

Non-syndromic

CISD2 CDGSH iron sulfur domain 2 Non-syndromicCLCNKB Chloride channel, voltage-sensitive

KbSyndromic

GJA1 Connexin 43 Non-syndromicGJB6 Connexin 30 Non-syndromicKCNQ1 Voltage-gated potassium channel SyndromicMYH9 Myosin, heavy chain 9, non-muscle Non-syndromicMYO6 Myosin 6 Non-syndromicOTOA Otoancorin Non-syndromicSLC12A2 Basolateral Na+-K+-Cl− symporter SyndromicTECTA Tectorin alpha Non-syndromic

190 J O U R N A L O F P R O T E O M I C S 1 0 3 ( 2 0 1 4 ) 1 7 8 – 1 9 3

[60], 380 in kidney epithelial cells [61], and 216 from neonatalmouse brain [62]. The detection of Cav1 in our screen providesinternal validation that DRM fractions contained sufficientquantities of DRM markers identified by qPCR, immunohisto-chemistry, and immunoblot. However, Cav3 was absent fromthe LC–MS/MS results as was the BK-channel α subunitpreviously reported in chick cochlear DRM [18]. These omissionssuggest a detection limit in the proteomic analysis and themissed identification of some verifiable DRM proteins.

Our informatic predictions estimated that only about 40%of identified DRM proteins have transmembrane helices. How-ever, post-translational modifications including palmitoylation,myristoylation, and glypiation anchor proteins to raft-likedomains [12,47]. Predictive algorithms suggested that thesemodifications could account for the targeting of the majority ofnon-integral membrane proteins to DRM fractions. Additionally,DRM are rich in protein complexes, which may also anchornon-integral membrane proteins to membrane microdomains.With this under consideration, only 7% of identified proteinswere predicted to have neither transmembrane domains norany of these post-translational modifications and were notwell-characterized members of protein complexes. Interesting-ly, proteins in this minority group include prohibitin andapolipoproteins, which are purported lipid-raft constituents[63,64], suggesting that the prediction and GO analyses mayunderestimate the number of proteins that may in fact bemembrane, and specifically DRM, associated. It is especiallyinteresting to note that these post-translational modificationsare highly regulated and modulated processes, enabling the cellto exert another level of control over DRM incorporation.Palmitoylation is highly dynamic. Over its lifetime, a proteincan be palmitoylated and depalmitoylated several timesallowing for dynamic incorporation and exclusion from micro-domains [47]. For example palmitoylation of inactive H-Rastargets the protein from Golgi to the plasma membrane, anddepalmitoylation results in H-Ras internalization [4]. H-Ras isinvolved in gentamicin-related ototoxicity [65], and its presencein cochlear DRM could provide novel links between fatty-acidmodifications, signaling compartmentalization, and hearingloss due to aminoglycosides. Similarly, myristoylation andglypiation can participate in domain targeting [12] and can

operate in a regulated manner [66,67]. Taken together, theprevalence of cochlear DRM proteins with putative fatty acidmodifications highlights the possibility that lipid-organizedplatforms in the ear are highly dynamic in nature andcontribute to a large array of regulatory processes.

The impact of cytoskeletal and organelle-associated pro-teins in DRM fractions is poorly understood. Indeed, theiridentification as raft constituents remains controversial.However, cytoskeletal proteins are known to interact withmany well-known raft proteins, and such interactions may bevital for the stabilization and function of the microdomain[12,68]. As an example, proteins involved in vesicle transportand actin-myosin motility were enriched in our DRM frac-tions. As a molecular motor, myosin is an essential compo-nent of membrane recycling, transport, and remodeling [69].Some myosins have been implicated in the organization ofcholesterol-rich membranes [70]. In hair cells specifically,myosin 1c, the hair cell's adaptationmotor, exhibits non-ionicdetergent resistance [71,72]. The presence of myosin 1c in ourcholesterol-rich DRM supports the hypothesis that elementsof the transduction apparatus rely on lipid-mediated com-partmentalization [20,73].

The enrichment of endoplasmic reticulum, Golgi, andmitochondrial proteins in DRM is even less well understoodas these membranes contain little cholesterol [12]. In manycases, these proteins may be recruited to the cell surface[74–77]. On the other hand, several lines of evidence suggestthat intracellular organelles contain raft-like domains. Raft-dependent endocytosis [78,79] and Golgi membrane cycling[80] indicate that lipid microdomains are involved in proteinsorting and vesicular trafficking between organelles andwith the cell surface. Similarly, raft-associated lipids includ-ing glycosphingolipids and polysialogangliosides are presentin mitochondria [12,81] and raft-associated proteins likeprohibitin and erlin have been identified in mitochondrialmembrane and ER membrane, respectively [63]. Thesefindings support the possibility that intracellular organellescontain raft-like microdomains that differ in compositionfrom plasma membrane microdomains and require lesscholesterol [63].

Two major themes emerge from the bioinformatic analysis,namely DRM involvement in metabolism and ion homeostasis.The large number of proteins associated with bioenergetics andmetabolism included several respiratory chain and mitochon-drial transport proteins, and the product of the non-syndromicdeafness gene Cisd2. Oxidative stress, mediated by Rac/Rhopathways and dysfunctional energetics, is commonly associat-ed with antibiotic-related ototoxicity, noise trauma, and age-related hearing loss [82–84]. The prevalence of metabolism-related proteins in inner ear DRM, as well as both Rac1 andRhoA, suggests a possible link between DRM organization andoxidative stress.

In addition to bioenergetics, proteins involved in ionhomeostasis were major components of the cochlear DRM.The cochlea is comprised of three fluid filled chambers anddifferences in the ionic composition of these fluids as well as anassociated electrical potential between the fluids is a criticalfeature of normal cochlear function. A hallmark of ionhomeostasis in the ear is the recycling of potassium throughthese fluid compartments. Cochlear DRM included the major

191J O U R N A L O F P R O T E O M I C S 1 0 3 ( 2 0 1 4 ) 1 7 8 – 1 9 3

ion channel responsible for potassium secretion into endo-lymph (KCNQ1) [85], gap junction proteins essential for the fluxof potassium through cellular compartments (GJA1 and GJB6)[86,87], the co-transporter largely responsible for establishingthe endocochlear potential (SLC12A2) [88], and tight-junctionproteins that facilitate separation of the endolymphatic–perilymphatic compartments.

Our results provide new insights into the possible impact ofcholesterol-rich microdomains in the ear and provide newhypotheses for uncovering the unique sensitivity of the ear tocholesterol-chelating cyclodextrins. In cats and mice, systemicapplication of 2-hydroxypropyl-beta-cyclodextrin causes sub-stantial outer hair cell death and hearing loss with no obviouspathology in other systems [27,89]. The unique sensitivity of theear to cyclodextrins is perplexing given the ubiquitous distribu-tion of cholesterol in cells throughout the body. However, theear is particularly sensitive to oxidative stress and ion imbal-ance, two processes heavily linked to DRM composition. Thesedata add to prior reports of cyclodextrin effects on cochlearmechanics [22,23,25] and the electrical activity of hair cells,specifically calcium influx and potassium efflux [18]. The abilityof cyclodextrins to function as a cholesterol shuttle has beenutilized as a treatment for Neiman–Pick C disease, a usually fataldisorder, where NPC1 or NPC2 deficiency prevents cholesterolefflux to the plasma membrane [90]. Hearing loss is part of theclinical diagnosis and spectrum of this disease and is associatedwith other genetic cholesterol synthesis disorders [91–94]. Theprofound sensitivity of the auditory organ to cyclodextrinsimplores further study into the source of cholesterol synthesisin the inner ear and the relationship between cholesterolhomeostasis and hearing loss.

5. Conclusion

The presence of over 600 proteins in inner ear cholesterol-enrichedmembranemicrodomains along with the prevalenceof microdomains in a variety of inner ear cell types indicatesthe importance of lipid-mediated compartmentalization inthis organ. Increasing evidence of lipid and cholesterolsegregation in the hair cell and the unique sensitivity of theauditory organ to changes in cholesterol localization motivatefurther study into the mechanism of cholesterol synthesisand regulation in this system.

Supplementary data to this article can be found online athttp://dx.doi.org/10.1016/j.jprot.2014.03.037.

Transparency document

The Transparency document associated with this article canbe found, in the online version.

Acknowledgments

This work was supported by grants from the NationalInstitutes of Health (R01 DC07432 to RKD and core grant P30DC05188) and Action on Hearing Loss (UK).

R E F E R E N C E S

[1] Krogh A, Larsson B, von Heijne G, Sonnhammer EL. Predictingtransmembrane protein topology with a hidden Markov model:application to complete genomes. J Mol Biol 2001;305:567–80.

[2] Pierleoni A, Martelli PL, Casadio R. PredGPI: a GPI-anchorpredictor. BMC Bioinformatics 2008;9:392.

[3] Maurer-Stroh S, Eisenhaber B, Eisenhaber F. N-terminalN-myristoylation of proteins: prediction of substrate proteinsfrom amino acid sequence. J Mol Biol 2002;317:541–57.

[4] Resh MD. Trafficking and signaling by fatty-acylated andprenylated proteins. Nat Chem Biol 2006;2:584–90.

[5] Singer SJ, Nicolson GL. The fluid mosaic model of thestructure of cell membranes. Science 1972;175:720–31.

[6] Engelman DM. Membranes are more mosaic than fluid.Nature 2005;438:578–80.

[7] Kusumi A, Ike H, Nakada C, Murase K, Fujiwara T.Single-molecule tracking of membrane molecules: plasmamembrane compartmentalization and dynamic assembly ofraft-philic signaling molecules. Semin Immunol 2005;17:3–21.

[8] Kusumi A, Nakada C, Ritchie K, Murase K, Suzuki K,Murakoshi H, et al. Paradigm shift of the plasma membraneconcept from the two-dimensional continuum fluid to thepartitioned fluid: high-speed single-molecule tracking ofmembrane molecules. Annu Rev Biophys Biomol Struct2005;34:351–78.

[9] Pike LJ. Lipid rafts: heterogeneity on the high seas. Biochem J2004;378:281–92.

[10] Lingwood D, Simons K. Lipid rafts as a membrane-organizingprinciple. Science 2010;327:46–50.

[11] Pike LJ. Rafts defined: a report on the Keystone Symposiumon Lipid Rafts and Cell Function. J Lipid Res 2006;47:1597–8.

[12] Pike LJ. The challenge of lipid rafts. J Lipid Res 2009;50:S323–8.[13] Karnovsky MJ, Kleinfeld AM, Hoover RL, Klausner RD. The

concept of lipid domains in membranes. J Cell Biol1982;94:1–6.

[14] Brown DA, Rose JK. Sorting of GPI-anchored proteins toglycolipid-enriched membrane subdomains during transportto the apical cell surface. Cell 1992;68:533–44.

[15] Head BP, Insel PA. Do caveolins regulate cells by actionsoutside of caveolae? Trends Cell Biol 2007;17:51–7.

[16] Saito K. Fine structure of the sensory epithelium ofguinea-pig organ of Corti: subsurface cisternae and lamellarbodies in the outer hair cells. Cell Tissue Res 1983;229:467–81.

[17] Saito K. Fine structure of the sensory epithelium of theguinea pig organ of Corti: afferent and efferent synapses ofhair cells. J Ultrastruct Res 1980;71:222–32.

[18] Purcell EK, Liu L, Thomas PV, Duncan RK. Cholesterolinfluences voltage-gated calcium channels and BK-typepotassium channels in auditory hair cells. PLoS One 2011;6:e26289.

[19] Nguyen T-VN, Brownell WE. Contribution of membranecholesterol to outer hair cell lateral wall stiffness.Otolaryngology—head and neck. Surgery 1998;119:14–20.

[20] Zhao H, Williams DE, Shin J-B, Brügger B, Gillespie PG. Largemembrane domains in hair bundles specify spatiallyconstricted radixin activation. J Neurosci 2012;32:4600–9.

[21] Sturm AK, Rajagopalan L, Yoo D, Brownell WE, Pereira FA.Functional expression and microdomain localization ofprestin in cultured cells. Otolaryngol Head Neck Surg2007;136:434–9.

[22] Sfondouris J, Rajagopalan L, Pereira FA, Brownell WE.Membrane composition modulates prestin-associated chargemovement. J Biol Chem 2008;283:22473–81.

[23] Rajagopalan L, Greeson JN, Xia A, Liu H, Sturm A, Raphael RM,et al. Tuning of the outer hair cell motor by membranecholesterol. J Biol Chem 2007;282:36659–70.

192 J O U R N A L O F P R O T E O M I C S 1 0 3 ( 2 0 1 4 ) 1 7 8 – 1 9 3

[24] Rajagopalan L, Organ-Darling L, Liu H, Davidson A, Raphael R,Brownell W, et al. Glycosylation regulates prestin cellularactivity. J Assoc Res Otolaryngol 2010;11:39–51.

[25] Brownell WE, Jacob S, Hakizimana P, Ulfendahl M, FridbergerA. Membrane cholesterol modulates cochlearelectromechanics. Pflugers Arch 2011;461:677–86.

[26] Levic S, Yamoah EN. Plasticity in membrane cholesterolcontributes toward electrical maturation of hearing. J BiolChem 2011;286:5768–73.

[27] Crumling MA, Liu L, Thomas PV, Benson J, Kanicki A, KabaraL, et al. Hearing Loss and hair cell death in mice given thecholesterol-chelating agent hydroxypropyl-β-cyclodextrin.PLoS One 2012;7:e53280.

[28] Li Y, Atkin G, Morales M, Liu L, Tong M, Duncan RK.Developmental expression of BK channels in chick cochlearhair cells. BMC Dev Biol 2009;9:67.

[29] Kim JM, Beyer R, Morales M, Chen S, Liu LQ, Duncan RK.Expression of BK‐type calcium‐activated potassium channelsplice variants during chick cochlear development. J CompNeurol 2010;518:2554–69.

[30] Paulo JA, Urrutia R, Banks PA, Conwell DL, Steen H. Proteomicanalysis of an immortalized mouse pancreatic stellate cellline identifies differentially-expressed proteins in activatedvs nonproliferating cell states. J Proteome Res2011;10:4835–44.

[31] Zybailov B, Mosley AL, Sardiu ME, Coleman MK, Florens L,Washburn MP. Statistical analysis of membrane proteomeexpression changes in Saccharomyces cerevisiae. J ProteomeRes 2006;5:2339–47.

[32] Götz S, García-Gómez JM, Terol J, Williams TD, Nagaraj SH,Nueda MJ, et al. High-throughput functional annotation anddata mining with the Blast2GO suite. Nucleic Acids Res2008;36:3420–35.

[33] Conesa A, Götz S, García-Gómez JM, Terol J, Talón M, RoblesM. Blast2GO: a universal tool for annotation, visualizationand analysis in functional genomics research. Bioinformatics2005;21:3674–6.

[34] Conesa A, Götz S. Blast2Go: a comprehensive suite forfunctional analysis in plant genomics. Int J Plant Genomics2008;2008:12.

[35] Götz S, Arnold R, Sebastián-León P, Martín-Rodríguez S,Tischler P, Jehl M-A, et al. B2G-FAR, a species-centered GOannotation repository. Bioinformatics 2011;27:919–24.

[36] Bindea G, Mlecnik B, Hackl H, Charoentong P, Tosolini M,Kirilovsky A, et al. ClueGO: a Cytoscape plug-in to decipherfunctionally grouped gene ontology and pathway annotationnetworks. Bioinformatics 2009;25:1091–3.

[37] Smoot ME, Ono K, Ruscheinski J, Wang PL, Ideker T.Cytoscape 2.8: new features for data integration and networkvisualization. Bioinformatics 2011;27:431–2.

[38] Rivals I, Personnaz L, Taing L, Potier M-C. Enrichment ordepletion of a GO category within a class of genes: whichtest? Bioinformatics 2007;23:401–7.

[39] Sonnhammer EL, von Heijne G, Krogh A. A hidden Markovmodel for predicting transmembrane helices inprotein sequences. Proc Int Conf Intell Syst Mol Biol1998;6:175–82.

[40] Ren J, Wen L, Gao X, Jin C, Xue Y, Yao X. CSS-Palm 2.0: anupdated software for palmitoylation sites prediction. ProteinEng Des Sel 2008;21:639–44.

[41] Xue Y, Liu Z, Cao J, Ren J. Computational prediction ofpost-translational modification sites in proteins. Ning-SunYang; 2011.

[42] Bologna G, Yvon C, Duvaud S, Veuthey AL. N‐terminalmyristoylation predictions by ensembles of neural networks.Proteomics 2004;4:1626–32.

[43] Fischer FP. Hair cell morphology and innervation in thebasilar papilla of the emu (Dromaius novaehollandiae). Hear Res1998;121:112–24.

[44] Selvaraj V, Asano A, Buttke DE, Sengupta P, Weiss RS, TravisAJ. Mechanisms underlying the micron-scale segregation ofsterols and GM1 in live mammalian sperm. J Cell Physiol2009;218:522–36.

[45] Blank N, Schiller M, Krienke S, Wabnitz G, Ho AD, Lorenz HM.Cholera toxin binds to lipid rafts but has a limited specificityfor ganglioside GM1. Immunol Cell Biol 2007;85:378–82.

[46] Henke RC, Hancox KA, Jeffrey PL. Characterization of twodistinct populations of detergent resistant membranecomplexes isolated from chick brain tissues. J Neurosci Res1996;45:617–30.

[47] Smotrys JE, Linder ME. Palmitoylation of intracellularsignaling proteins: regulation and function. Annu RevBiochem 2004;73:559–87.

[48] Zhao HB, Kikuchi T, Ngezahayo A, White TW. Gap junctionsand cochlear homeostasis. J Membr Biol 2006;209:177–86.

[49] Scherer PE, Lewis RY, Volonte D, Engelman JA, Galbiati F, Couet J,et al. Cell-type and tissue-specific expression of caveolin-2.Caveolins 1 and 2 co-localize and form a stablehetero-oligomeric complex in vivo. J Biol Chem1997;272:29337–46.

[50] Head BP, Hu Y, Finley JC, Saldana MD, Bonds JA, MiyanoharaA, et al. Neuron-targeted caveolin-1 protein enhancessignaling and promotes arborization of primary neurons. JBiol Chem 2011;286:33310–21.

[51] Okamoto T, Schlegel A, Scherer PE, Lisanti MP. Caveolins, afamily of scaffolding proteins for organizing "preassembledsignaling complexes" at the plasma membrane. J Biol Chem1998;273:5419–22.

[52] Li X, McClellan ME, Tanito M, Garteiser P, Towner R, Bissig D,et al. Loss of caveolin-1 impairs retinal function due todisturbance of subretinal microenvironment. J Biol Chem2012;287:16424–34.

[53] Mikol DD, Hong HL, Cheng HL, Feldman EL. Caveolin-1expression in Schwann cells. Glia 1999;27:39–52.

[54] Pang H, Le PU, Nabi IR. Ganglioside GM1 levels are adeterminant of the extent of caveolae/raft-dependentendocytosis of cholera toxin to the Golgi apparatus. J Cell Sci2004;117:1421–30.

[55] Tucker T, Fettiplace R. Confocal imaging of calciummicrodomains and calcium extrusion in turtle hair cells.Neuron 1995;15:1323–35.

[56] Martinez-Dunst C, Michaels RL, Fuchs PA. Release sites andcalcium channels in hair cells of the chick's cochlea.J Neurosci 1997;17:9133–44.

[57] Li Y, Atkin GM, Morales MM, Liu LQ, Tong M, Duncan RK.Developmental expression of BK channels in chick cochlearhair cells. BMC Dev Biol 2009;9:67.

[58] Ostermeyer AG, Beckrich BT, Ivarson KA, Grove KE, BrownDA. Glycosphingolipids are not essential for formation ofdetergent-resistant membrane rafts in melanoma cells.methyl-beta-cyclodextrin does not affect cell surfacetransport of a GPI-anchored protein. J Biol Chem1999;274:34459–66.

[59] Brady JD, Rich TC, Le X, Stafford K, Fowler CJ, Lynch L, et al.Functional role of lipid raft microdomains in cyclicnucleotide-gated channel activation. Mol Pharmacol2004;65:503–11.

[60] von Haller PD, Donohoe S, Goodlett DR, Aebersold R,Watts JD.Mass spectrometric characterization of proteins extractedfrom Jurkat T cell detergent-resistant membrane domains.Proteomics 2001;1:1010–21.

[61] Blonder J, Hale ML, Lucas DA, Schaefer CF, Yu LR, Conrads TP,et al. Proteomic analysis of detergent-resistant membranerafts. Electrophoresis 2004;25:1307–18.

[62] Yu H, Wakim B, Li M, Halligan B, Tint GS, Patel SB.Quantifying raft proteins in neonatal mouse brain by‘tube-gel’ protein digestion label-free shotgun proteomics.Proteome Sci 2007;5:17.

193J O U R N A L O F P R O T E O M I C S 1 0 3 ( 2 0 1 4 ) 1 7 8 – 1 9 3

[63] Browman DT, Hoegg MB, Robbins SM. The SPFHdomain-containing proteins: more than lipid raft markers.Trends Cell Biol 2007;17:394–402.

[64] Browman DT, Resek ME, Zajchowski LD, Robbins SM. Erlin-1and erlin-2 are novel members of the prohibitin family ofproteins that define lipid-raft-like domains of the ER. J Cell Sci2006;119:3149–60.

[65] Battaglia A, Pak K, Brors D, Bodmer D, Frangos JA, Ryan AF.Involvement of ras activation in toxic hair cell damage of themammalian cochlea. Neuroscience 2003;122:1025–35.

[66] Ames JB, Lim S, Ikura M. Molecular structure and targetrecognition of neuronal calcium sensor proteins. Front MolNeurosci 2012;5:10.

[67] Mayor S, Riezman H. Sorting GPI-anchored proteins. Nat RevMol Cell Biol 2004;5:110–20.

[68] Oliferenko S, Paiha K, Harder T, Gerke V, Schwärzler C,Schwarz H, et al. Analysis of Cd44-containing lipid raftsrecruitment of annexin II and stabilization by the actincytoskeleton. J Cell Biol 1999;146:843–54.

[69] Diz-Munoz A, Fletcher DA, Weiner OD. Use the force:membrane tension as an organizer of cell shape and motility.Trends Cell Biol 2013;23:47–53.

[70] Hanzal-Bayer MF, Hancock JF. Lipid rafts and membranetraffic. FEBS Lett 2007;581:2098–104.

[71] Gillespie PG, Cyr JL. Myosin-1c, the hair cell's adaptationmotor. Annu Rev Physiol 2004;66:521–45.

[72] Gillespie PG, Wagner MC, Hudspeth A. Identification of a120 kD hair-bundle myosin located near stereociliary tips.Neuron 1993;11:581–94.

[73] Hirono M, Denis CS, Richardson GP, Gillespie PG. Hair cellsrequire phosphatidylinositol 4,5-bisphosphate formechanical transduction and adaptation. Neuron2004;44:309–20.

[74] Kim KB, Lee JW, Lee CS, Kim BW, Choo HJ, Jung SY, et al.Oxidation–reduction respiratory chains and ATP synthasecomplex are localized in detergent‐resistant lipid rafts.Proteomics 2006;6:2444–53.

[75] Kim BW, Lee JW, Choo HJ, Lee CS, Jung SY, Yi JS, et al.Mitochondrial oxidative phosphorylation system is recruitedto detergent‐resistant lipid rafts during myogenesis.Proteomics 2010;10:2498–515.

[76] Bini L, Pacini S, Liberatori S, Valensin S, Pellegrini M,Raggiaschi R, et al. Extensive temporally regulatedreorganization of the lipid raft proteome following T-cellantigen receptor triggering. Biochem J 2003;369:301–9.

[77] Bae T-J, Kim M-S, Kim J-W, Kim B-W, Choo H-J, Lee J-W, et al.Lipid raft proteome reveals ATP synthase complex in the cellsurface. Proteomics 2004;4:3536–48.

[78] Helms JB, Zurzolo C. Lipids as targeting signals: lipid rafts andintracellular trafficking. Traffic 2004;5:247–54.

[79] Ikonen E. Roles of lipid rafts in membrane transport. CurrOpin Cell Biol 2001;13:470–7.

[80] Nichols BJ, Kenworthy AK, Polishchuk RS, Lodge R, RobertsTH, Hirschberg K, et al. Rapid cycling of lipid raft markersbetween the cell surface and Golgi complex. J Cell Biol2001;153:529–42.

[81] Gillard BK, Thurmon LT, Marcus DM. Variable subcellularlocalization of glycosphingolipids. Glycobiology 1993;3:57–67.

[82] Jiang H, Sha SH, Schacht J. Rac/Rho pathway regulates actindepolymerization induced by aminoglycoside antibiotics.J Neurosci Res 2006;83:1544–51.

[83] Chen F-Q, Zheng H-W, Hill K, Sha S-H. Traumatic noiseactivates Rho-family GTPases through transient cellularenergy depletion. J Neurosci 2012;32:12421–30.

[84] Jiang H, Talaska AE, Schacht J, Sha SH. Oxidative imbalance inthe aging inner ear. Neurobiol Aging 2007;28:1605–12.

[85] Wangemann P. K+ cycling and the endocochlear potential.Hear Res 2002;165:1–9.

[86] Liu XZ, Xia XJ, Adams J, Chen ZY, Welch KO, Tekin M, et al.Mutations in GJA1 (connexin 43) are associated withnon-syndromic autosomal recessive deafness. Hum MolGenet 2001;10:2945–51.

[87] Teubner B, Michel V, Pesch J, Lautermann J, Cohen-SalmonM,Söhl G, et al. Connexin30 (Gjb6)-deficiency causes severehearing impairment and lack of endocochlear potential. HumMol Genet 2003;12:13–21.

[88] Dixon MJ, Gazzard J, Chaudhry SS, Sampson N, Schulte BA,Steel KP. Mutation of the Na-K-Cl co-transporter gene Slc12a2results in deafness in mice. Hum Mol Genet 1999;8:1579–84.

[89] Ward S, O'Donnell P, Fernandez S, Vite CH.2-Hydroxypropyl-[bgr]-cyclodextrin raises hearing thresholdin normal cats and in cats with Niemann–Pick type C disease.Pediatr Res 2010;68:52–6.

[90] Rosenbaum AI, Maxfield FR. Niemann–Pick type C disease:molecular mechanisms and potential therapeutic ap-proaches. J Neurochem 2011;116:789–95.

[91] Di Berardino F, Alpini D, Ambrosetti U, Amadeo C, Cesarani A.Sensorineural hearing-loss in the Smith–Lemli–Opitzsyndrome. Int J Pediatr Otorhinolaryngol Extra 2007;2:169–72.

[92] McCall AA, Kirsch CF, Ishiyama G, Ishiyama A. Otologicfindings in Antley–Bixler syndrome: a clinical and radiologiccase report. Int J Pediatr Otorhinolaryngol 2007;71:1139–43.

[93] Yanjanin NM, Velez JI, Gropman A, King K, Bianconi SE,Conley SK, et al. Linear clinical progression, independent ofage of onset, in Niemann–Pick disease, type C. Am J MedGenet B Neuropsychiatr Genet 2010;153B:132–40.

[94] Brunetti-Pierri N, Corso G, Rossi M, Ferrari P, Balli F, Rivasi F,et al. Lathosterolosis, a novel multiple-malformation/mentalretardation syndrome due to deficiency of3beta-hydroxysteroid-delta5-desaturase. Am J Hum Genet2002;71:952–8.